Dear all,

This refers to a situation I faced in containing vibration in flash tank in a power plant a decade ago. We had a control valve with upstream & downstream isolation & bypass globe valve dischaging high velocity mixed phase fluid in a flash tank. The flash tank was connected to condenser & was vibrating violently, when plant operation started. I suggested joining of bypass valve to downstream header directly through an expander and control valve joined to the tangential entry header through an expander, deleting all elbows downstream control valve & bypass globe valve. We had to strengthen the flash vessel supports against the torsional effect the vessel was subjected to due to tangential entry by welding with support structure. The vibration reduced, although every 10 years, the flash vessels needed replacement due to erosion. We solved the problem, but not analysed.

Now, my question is, if the vessel was vibrating in tangential direction by 10 mm observed amplitude, how would we predict the fatigue life of the vessel ? Is it possible to simulate the situation in caesar-II by introducing harmonic displacements ? With alternating stress evaluated, using matrial specific appropriate endurance test curve & guess estimate of fequency of vibration from signature study made at flash tank incoming header,can the number of cycles of safe operation of flash tank be predicted ? If already the plant had been subjected to this vibration unchecked for say 1000 cycles & the predicted cycle life is 10000 cycles, can it remain safe for 9000 cycles ?

regards,

sambhu

I agree with your suggestion for a Harmonic Analysis.But make sure you specify the HARMONIC DISPLACEMENTS at the correct point in your model. To this question , my suggestion is : specify the displacement at Pipe -Nozzle connection.The reason for doing so is : the Flash Drum is stronger w.r.t the Pipe and it is the Flash Drum which is making the Piping vibrate.The input to displacement in Harmonic analysis is the "Imposed displacement" and the resulting displacement at other nodes of the system are the "response displacement " which the program has to compute.I have seen many stress engineers specify the field measured displacement values at other nodes as input in their model , which is incorrect.

The ASME sec VIII DIv 2 Appendix 5 fatigue curves are for polished specimens and should be used with caution for Piping systems.

The Harmic analysis in Caesar II Is an apt method to predict the fatigue life.

You can go for another method. Check the field measurements for displament/velocity and check against the allowable displacements/velocity. There are numerous published papers on this subject of allowable data for field vibration measurements.


Anindya Bhattacharya

The flowing reference may help:

Bulletin 417 December 1996
Welding Research Council
Design Guide to Reduce Potential For vibration Caused by Fluid Flow Inside Pipes-Review and Survey
C.W. LIN
ISSN 0043-2326

www.thill.biz